Outcomes: Students gain skills in evaluating, experimenting with, and optimising the performance of computer architectures from single processors to parallel computers; understand the foundational concepts needed for learning about computer systems and describing the technological context of current computer organisations; and understand the foundation concepts of various classes of parallel computers, which is expanded to evaluating and optimising their performance.

Content: Topics include components of a computer, classification of computer architectures, Von Neumann and Harvard architectures, Flynn’s classification, benchmarking and performance measurement, instruction set architecture, CPU components, CPU internal buses, microprogramming, hardwired control, input/output (I/O), bus organisation, DMA I/O, memory-mapped I/O, memory systems, caches, virtual memory systems, memory speed and cost-performance issues, pipelining and RISCs, processor design and performance issues. Introduction to parallel computing, requirements for parallelism: data dependencies, precedence graphs, measuring parallelism: level, degree and granularity of parallelism, measurement of performance, speed-up laws, MIMD multiprocessing, SIMD computing, multi-pipeline systems and pipeline chaining, control and scheduling of arithmetic and instruction pipelines; inter-processor connection networks.